Prognostic value of end-to-end deep learning assessment of myocardial scar and microvascular obstruction on late gadolinium enhancement cardiovascular magnetic resonance.
Authors
Affiliations (12)
Affiliations (12)
- Department of Radiology, The Second Affiliated Hospital, Jiangxi Medical College, Nanchang University, Nanchang, China; Jiangxi Provincial Key Laboratory of Intelligent Medical Imaging, Nanchang, China; National Heart Research Institute Singapore, National Heart Centre Singapore, Singapore.
- National Heart Research Institute Singapore, National Heart Centre Singapore, Singapore; Cardiovascular Sciences Academic Clinical Programme, Duke-NUS Medical School, Singapore.
- National Heart Research Institute Singapore, National Heart Centre Singapore, Singapore.
- Cardiovascular Sciences Academic Clinical Programme, Duke-NUS Medical School, Singapore; Department of Cardiology, National Heart Centre Singapore, Singapore.
- Department of Radiology, The Second Affiliated Hospital, Jiangxi Medical College, Nanchang University, Nanchang, China; Jiangxi Provincial Key Laboratory of Intelligent Medical Imaging, Nanchang, China.
- Department of Cardiology, National University Heart Centre, Singapore; Yong Loo Lin School of Medicine, National University of Singapore, Singapore.
- Yong Loo Lin School of Medicine, National University of Singapore, Singapore; Department of Diagnostic Imaging, National University Hospital, Singapore.
- Mayo Clinic in Rochester, Rochester, Minnesota, United States.
- Cardiovascular Sciences Academic Clinical Programme, Duke-NUS Medical School, Singapore; Department of Cardiology, National Heart Centre Singapore, Singapore; Cardiovascular & Metabolic Disorders Program, Duke-NUS Medical School, Singapore.
- Department of Radiology, The Second Affiliated Hospital, Jiangxi Medical College, Nanchang University, Nanchang, China; Jiangxi Provincial Key Laboratory of Intelligent Medical Imaging, Nanchang, China. Electronic address: [email protected].
- National Heart Research Institute Singapore, National Heart Centre Singapore, Singapore; Cardiovascular Sciences Academic Clinical Programme, Duke-NUS Medical School, Singapore; Yong Loo Lin School of Medicine, National University of Singapore, Singapore; Cardiovascular & Metabolic Disorders Program, Duke-NUS Medical School, Singapore; The Hatter Cardiovascular Institute, University College London, London, UK.
- National Heart Research Institute Singapore, National Heart Centre Singapore, Singapore; Cardiovascular Sciences Academic Clinical Programme, Duke-NUS Medical School, Singapore; Cardiovascular & Metabolic Disorders Program, Duke-NUS Medical School, Singapore; Department of Biomedical Engineering, National University of Singapore, Singapore. Electronic address: [email protected].
Abstract
Late gadolinium enhancement (LGE) cardiovascular magnetic resonance (CMR) is the reference standard for assessing myocardial scar and microvascular obstruction (MVO), strong predictors of post-acute myocardial infarction (AMI) outcomes. However, manual segmentation is time-consuming and subject to inter-observer variability, limiting clinical scalability. This study develops and validates LGE-CMRnet, an end-to-end deep learning pipeline for automated scar and MVO segmentation on LGE CMR, and evaluates its prognostic value in AMI patients. A total of 3,874 LGE images from 567 AMI patients (409 for training/internal stress-test cohort; 158 for external testing) were analyzed. LGE-CMRnet integrates YOLOv8 for heart localization and nnU-Net for simultaneous segmentation of myocardium, scar, and MVO. Performance was evaluated using Dice similarity coefficient (DSC), correlation, and Bland-Altman analysis against expert annotations. Prognostic value was assessed using Cox regression for major adverse cardiac events (MACE) over a median follow-up of 24.4 months. LGE-CMRnet achieved rapid processing (0.05seconds per image) and high segmentation accuracy. In the external validation cohort, the model achieved mean DSC of 0.83±0.11 for scar and 0.88±0.11 for MVO at the patient level, with strong volumetric correlations to expert reference segmentations (scar: r=0.90; MVO: r=0.98, both P<0.0001). Bland-Altman analysis showed minimal bias in volumetric measurements (scar: 2.5±8.9 cm<sup>3</sup>; MVO: -0.20±0.89 cm<sup>3</sup>). Among the 158 patients in the external validation cohort (age 57±10 years, 80% male), 35 (22.2%) experienced MACE. LGE-CMRnet-derived %MVO (hazard ratio [HR], 1.06; 95% confidence interval [CI]: 1.02 to 1.09; P=0.003) and %Scar (HR, 1.05; 95% CI: 1.02 to 1.08; P=0.001) were independent predictors of MACE after adjustment for established risk factors. Furthermore, LGE-CMRnet-derived metrics demonstrated non-inferior discrimination for MACE prediction compared with expert analysis. The differences in C-index were 0.02 for %MVO and 0.01 for %Scar, with the lower bounds of the 95% CIs remaining above the pre-specified non-inferiority margin. LGE-CMRnet enables fast and accurate scar and MVO quantification, with prognostic performance comparable to expert analysis, supporting its potential for automated clinical risk stratification after AMI.